Hi Ian,As you said, increasing people are shifting to Python 3 but a lot of people are still using Python 2.7 for now and many python library didn't shift yet so IMO, we should not switch to fast between Python 2.7 and Python 3.X Also, Python representation for Python 2.7 and Python 3.x are different so it would mean we need to maintains a lots more code for this support and I am not sure we have enough engineering time for this (but if someone want to contribute it, let's go !!! ) For these reasons, I think Python 3.X will not be available soon in Pythran but two step can be done. 1) switching internal Python code from 2.7 to 3.X (or having a code supporting both versions) 2) having a switch to generate Python 3.X pythran modules from Python 2.7 code.
Once these steps are done, a total switch from 2.7 to 3.x could be considers.
Serge may have another opinion. Pierrick On 06/11/2014 15:34, Ian Ozsvald wrote:
Hi all. I taught a 2 day High Performance Python course last week, I've upgraded all my material to Python 3.4. I talked about pythran along with giving demos of Numba, PyPy and focusing on Cython (these 3 all have Python 3+ support).Do you have plans to support Python 3? Within 5 years Python 2.7 will be dead (the sunset date is 2020) and I get the feeling that increasingly people are shifting up to Python 3, then not looking back. I use Python 3.4 for all client work now - sklearn, numpy, nltk, pymongo, pymysql etc all work fine using Anaconda.Ian. -- Ian Ozsvald (A.I. researcher) ian@xxxxxxxxxxxxxx http://IanOzsvald.com http://ModelInsight.io http://MorConsulting.com http://Annotate.IO http://SocialTiesApp.com http://TheScreencastingHandbook.com http://FivePoundApp.com http://twitter.com/IanOzsvald http://ShowMeDo.com